1. Field of the Invention
The present invention relates to glycopeptide compounds and libraries of glycopeptide compounds structurally analogous to known glycopeptide antibiotics and methods of generating those libraries. The compounds contain modified carbohydrate moieties. The libraries are generated using combinatorial chemical techniques that produce a diverse set of carbohydrate functionalities conjugated to an oligopeptide.
2. Background of the Invention
Glycopeptide antibiotics are characterized by having at least one saccharide group chemically bonded to a rigid peptide structure having a cavity or cleft which acts as a binding site for the substrate used in bacterial cell wall synthesis. The glycopeptide antibiotics are further categorized into various subclasses depending on the identity and interconnections of the amino acids comprising the peptide backbone and the number and substitution pattern of the sugar residues in the molecule. The glycopeptide antibiotics are generally active against Gram-positive bacteria but relatively ineffective against Gram-negative bacteria
Most notable among the glycopeptide antibiotics is vancomycin. Vancomycin is produced by Amycolatopsis orientalis, and is often referred to as “the drug of last resort” because it is effective against most multi-drug-resistant gram positive bacteria. However, in recent years vancomycin-resistant strains of some bacteria have emerged. [Cohen M., (1992); Neu H., (1992)]. It is estimated that 5-25% of enterococcal strains in hospitals are now resistant to vancomycin [Axelsen, P. H. et al. (1997)]. Most feared among the bacteria is Staphylococcus aureus, which can result in dangerous respiratory and blood infections. Vancomycin-resistant and vancomycin-insensitive strains of this bacterium have also been recently reported [Milewski (1996)].
The structural formula of vancomycin is shown below and is characterized by a disaccharide moiety covalendy linked to a heptapeptide structure. The structure of vancomycin places it in a class of molecules referred to as the “dalbaheptides.” [Malabarba A., et al. (1997a)] Dalbaheptides in general are characterized by the presence of seven amino acids linked together by peptide bonds and held in a rigid conformation by cross-links through the aromatic substituent groups of at least five of the amino acid residues. In the heptapeptide structure of vancomycin, which is commonly referred to as the “aglycone” of vancomycin, the aromatic side-chains of amino acids 2, 4, and 6 are fused together through ether linkages. The side-chains of amino acids 5 and 7 are joined via a carbon-carbon bond. Amino acids 1 and 3 arm leucine and asparaine, respectively. Other naturally-occurring glycopeptide antibiotics are similar to vancomycin in that they have a glucose residue linked to the aromatic substituent on amino acid 4 through formation of a bond with a phenolic hydroxyl group. The glucose residue, in turn, is linked through its vicinal hydroxyl position to a unique amino sugar, either L-vancosamine. The sugars have been separately removed from glycopeptide antibiotics, and it has been found that the presence of both sugars enhances the pharmacokinetic properties of this class of antibiotics. [Nagarajan R. (1988), (1991), (1993]

The anti-microbial activity of vancomycin is known to be due to its ability to interfere with biosynthesis of the bacterial cell wall. [Nagarajan R. (1993)]. NMR evidence shows that the heptapeptide chain of vancomycin forms a number of hydrogen bonds with the D-alanyl-D-alanine terminus of the disaccharide-pentapeptide precursors used to form the cell wall. [see, e.g., Prowse W., et al. (1995); Pierce C., et al. (1995); Williams D. et al. (1988)]. This interaction of vancomycin with cell wall precursors apparently inhibits or prevents the subsequent transglycosylation and/or transpeptidation steps of cell wall assembly. Supporting this mode of action is the fact that vancomycin-resistant strains of bacteria are found to produce a pentapeptide precursor terminating in a D-alanyl-D-lactate sequence. It is hypothesized that the reduced effectiveness of vancomycin against resistant strains is due to reduced hydrogen bonding interactions between the drug and the D-alanyl-D-lactate substrate. The affinity of vancomycin for D-alanyl-D-lactate is estimated to be 2-3 orders of magnitude (4.1 kcal/mol) less than for D-alanyl-D-alanine. [Walsh C. (1993)].
The sugar residues of the vancomycin and other glycopeptide antibiotics have been shown to affect binding activities. Structural changes in the sugar residues can produce significant changes in antibiotic activity. [Malabarba (1997), Nagarajan, R. (1993)] It has been proposed that the sugar residues on the glycopeptide antibiotics may enhance the avidity of these molecules for surface-bound peptide ligands. At least two different mechanisms for enhancing avidity have been proposed. [Kaiman (1988), Gerhard (1993), Allen (1997)]
For example, it has been proposed that the biological activity of vancomycin, along with that of many other glycopeptide antibiotics, is enhanced by dimerization due to bonding interactions at the convex (non-ligand binding) face of the molecule. [Williams D., et al. (1993); Gerhard U., et al., (1993)] Dimerization is believed to be facilitated by the disaccharide groups of the vancomycin molecule, and is thought to influence activity by increasing the avidity of vancomycin for surface-bound D-Ala-D-Ala peptide ligands. [Williams, (1998)] Structural evidence for dimerization has been obtained from both NMR and crystallographic studies, and it has been found that there are significant differences in the stability of the dimers formed in solution by different glycopeptide antibiotics. [MacKay (1994)] It is proposed that differences in the dimerization constants may account at least partially for the remarkable differences in biological activity of different glycopeptide antibiotics which otherwise have very similar binding affinities for the natural d-Ala-d-Ala substrate. [Williams (1998)]
A second mechanism for enhancing activity has also been proposed for the glycopeptide antibiotic teicoplanin, which contains an N-alkyl chain on one of the sugars. It is suggested that this N-alkyl chain increases the effective avidity of teicoplanin for surface-bound D-Ala-D-Ala ligands by interacting with the membrane, thus “anchoring” the teicoplanin molecule at the membrane surface. [Beures (1995)] It should be noted that the attachment of hydrophobic substituents to the vancomycin carbohydrate moiety appears to enhance activity against vancomycin-resistant strains. For example, attaching a hydrophobic group to the vancosamine sugar by alkylation on the amine nitrogen increases activity against vancomycin-resistant strains by two orders of magnitude. [Nagarajan (1991)] It is speculated that the lipophilic groups locate the antibiotic at the cell surface and make ligand binding an intramolecular process, which may partially overcome the decreased binding affinity for D-Ala-D-Lac. Hence, although the sugars on the glycopeptide antibiotics do not appear to interact substantially with the peptide substrates, they play a very important role in increasing the biological activity. Therefore, one potentially successful strategy for the design of new antibacterial agents based on the glycopeptide class of antibiotics involves modifying the carbohydrate portions of the molecules. [Malabarba (1997a)]
Related members of the vancomycin class of glycopeptide antibiotics include the ristocetins, the eremomycins, the avoparcins and teicoplanin. Several of these compounds are shown, together with vancomycin in FIGS. 1a and 1b. The chemical structures of all of these compounds include a dalbaheptide structure as the aglycone core, with minor differences in the amino acids and in cross-linking, but differ from each other most distinctively in terms of the nature of the sugar residues as well as the number and points of attachment of sugar residues to the aglycone core. It is known that biological activities of vancomycin-type antibiotics vary depending on the nature of the sugar residues.
One approach to obtaining new drug candidates derived from vancomycin and other glycopeptide antibiotics has involved chemical modification of one or more sugar residues of the naturally occurring glycopeptide. For instance, as noted previously, an aikyl chain can be attached to a sugar residue of the molecule, such as at the amino group of the amino sugar. [Cooper, R. et al. (1996)]. Other semi-synthetic approaches have involved traditional esterification and amidation methodologies applied to the peptide portion of the molecule. [Malabarba, A. et al. (1997b)] The attachment of lipophilic alkyl chains to the antibiotic has been proposed to afford better membrane anchoring, thereby increasing the effective activity of glycopeptide at the cell wall. [Felmingham, D. (1993)] The presence of an additional sugar has also produced compounds having enhanced activity, which may be due to their improved dimerization ability. [Malabarba A., et al. (1997a); Allen N. et al., (1997]. Other semi-synthetic approaches to modification of the vancomycin molecule have involved derivatization of the polypeptide binding pocket. [Pavlov A., et al. (1993)]
Previous efforts in producing new compounds having increased activity against vancomycin-resistant strains have typically involved a directed synthesis of a specific target derivative of a natural glycopeptide. This is a slow and relatively tedious process requiring a great deal of time and expense to obtain a suitable set of drug candidates for use in screening for activity. It is desirable to develop a combinatorial approach to the synthesis of new drug candidates based on the glycopeptide antibiotics. Recognizing this, Griffin and coworkers synthesized a combinatorial library of vancomycin derivatives in which different peptide chains were appended to the carboxylate on amino acid 7. No candidates were identified which had significantly improved activity compared with the underivatized natural product for either vancomycin-sensitive or vancomycin-resistant strains. The failure of the effort highlights a key requirement for a strategy involving the synthesis of a library related to a natural product: it is imperative to introduce substituents at positions on the molecule where there is evidence that such substitutions will have an effect on activity. In the case of the glycopeptide antibiotics, changes to the carbohydrate portions of the molecules would seem to be warranted in light of the relatively large role played by the sugar residues in increasing activity. The use of enzymes to generate glycosylated vancomycin derivatives wherein the saccharide residue carries a variety of functionalizations has been proposed and explored. [Solenberg (1997)] However, the range of compounds that can be prepared using enzymes in this manner is limited by the availability of enzymes specific to the desired functionalized saccharide residue. This has only been demonstrated for glucose and xylose; vancosamine has never been attached using the enzyme method and no compounds displaying activity have been produced using the enzyme method. No other strategies for making libraries of glycopeptide antibiotics in which the carbohydrate moieties are combinatorially varied have been reported.
Comparison of the natural products have made it clear that the nature and placement of the sugars on the glycopeptide antibiotics play critical roles in antibiotic activity. Furthermore, there is some information from semi-synthetic efforts about positions on the carbohydrates that may be important in activity. For example, we have already noted that some vancomycin derivatives containing hydrophobic substituents on the vancosamine nitrogen show improved activity against vancomycin-resistant strains. However, there have been no reports of modifications on the glucose residue of vancomycin which have affected activity. In fact, for glycopeptide antibiotics containing two or more sugars attached to amino acid 4, there is no suggestion in the literature that the sugar directly attached to the aglycone can be modified to improve activity. It has even been argued that the glucose residue “has no independent contribution to binding, and it is likely that its role with respect to the binding constant is merely to position the vancosamine optimally relative to the aglycon portion.” [Kannan et al. (1988)]
The structure-activity relationships among the vancomycin-like glycopeptide antibiotics show that the presence of an amino sugar at the residue 6 benzylic position and an N-alkyl or N-aryl substituted amino sugar at the amino acid-4 position increases antibiotic activity against both VRE and VSE. However, these trends do not always hold against other gram-positive bacteria such as the Staphylococci and Streptococci. Furthermore, no studies have addressed the effects of introducing functionality on the sugar groups other than N-alkylation, N-acylation, formation of N-oxides or modification of ester groups at C-6. Because the nature and placement of the sugars on glycopeptide antibiotics play such critical roles in antibiotic activity, many more studies are needed to optimize the sugar substituents. Such studies could not only lead to better antibiotics against vancomycin-resistant bacteria, but might provide more information about the mechanism of interaction at bacterial membranes. Preparation of derivatives with different sugar substituents will not only probe the sugar's role in currently proposed interactions, but may also lead to the discovery of new specific or non-specific interactions of the glycopeptide antibiotics at the cell surface. For reviews regarding the structure activity relationships of natural and semisynthetic glycopeptide antibiotics see Malabarba et al. Med. Res. Rev., 1997, 17, 69; Nagarajan, Antimicrob. Agents Chemother., 1991, 35, 605; Nagarajan, J. Antibiotics, 1993, 46, 1181; Cooper and Thompson, Ann. Rep. Med. Chem., 1996, 31, 131; Malabarba et al., Eur. J. Med. Chem., 1997, 32, 459; Allen et al. J. Antibiotics, 1997, 50, 677.
Combinatorial strategies have been successfully applied to the synthesis of peptide, nucleic acid, and various small molecule libraries, however, they have not been extensively employed to make carbohydrate-based libraries. Most of the approaches to production of carbohydrate libraries have been conducted in solution. A solid phase approach to making diverse libraries of di- and tri-saccharide compounds has also been reported. [Liang et al. (1996)]. A solid phase method permits reactions to be driven to completion by using a large excess of reactants. The solid phase approach also permits spatial resolution of the product compounds. Glycopeptide libraries have been produced on the solid phase in which amino acids were varied. However, no suggestion has been made that glycopeptide antibiotics can be made using a solid-phase method.